単一チャンネル睡眠時脳波による自動睡眠ステージ推定法の評価  [in Japanese] Sleep stage automatic classification with single channel sleep EEG signals  [in Japanese]

Author(s)

Abstract

本研究の目的は,単一チャンネル脳波による自動睡眠ステージ判定アルゴリズムが,臨床検査技師による終夜ポリソムノグラフィー(PSG)検査により推定された睡眠ステージとの比較評価をすることにある.脳波,眼振図,筋電図や呼吸センサによりなるPSG検査は病院施設における睡眠の質判定法としてよく用いられている.日常生活における長時間計測を可能にするために,本研究グループでは低電力,軽量な単一チャンネル睡眠時脳波計測器をすでに開発しており,それを用いた.14名の被験者から得られた睡眠ステージの自動推定をPSG検査との比較を行った結果,最大で81.9%の一致率であり,平均すると63.1%となった.専門技術者が単一チャンネル脳波により睡眠ステージを推定した結果,PSG検査との比較においても84%から95%程度であることから,現状のアルゴリズムについて推定精度が睡眠の質を十分に計測できたといえないまでも支援システムとしての機能を果たせることを示唆した.

The purpose of this study is to examine that automatic classification algorithm with the single channel EEG signal is compared with sleep stage classification with polysomnography (PSG) by the clinical laboratory technician. PSG which is composed of sleep EEG, EOG, EMG signals and respirometry, is frequently used to measure quality of sleep for fording sleep impairments in hospital. For the purpose of evaluate long term sleep condition in ordinary lifestyle, we have developed autonomic classification of sleep stage with just only single channel EEG signal which is low electric supply and light weight and easy-to-use. Our results show that the full automatic classification algorithm could have the corresponding rate 81.9% as the maximum and 63.1% as the totally average to 14 subject sleep EEG signals as compared with the classifications with PSG. It is suggested that the corresponding rate score with automatic one is higher because corresponding rate scores with a technician to single channel EEG signal are only at most about 95% and the average under 84%. However, several EEG signals with lower corresponding rates were also there. The additional standards, c.f. each of characteristic waveforms and predictive information between epochs to classify them, should be included in the algorithm. In conclusion, in most cases, our results show that our developed algorithms have the capability to support assessing sleep stages as a screening diagnosis for sleep quality.

Journal

  • IEICE technical report. ME and bio cybernetics

    IEICE technical report. ME and bio cybernetics 111(234), 33-36, 2011-10-06

    The Institute of Electronics, Information and Communication Engineers

References:  4

Codes

  • NII Article ID (NAID)
    110008900134
  • NII NACSIS-CAT ID (NCID)
    AN1001320X
  • Text Lang
    JPN
  • Article Type
    ART
  • ISSN
    0913-5685
  • NDL Article ID
    023340636
  • NDL Call No.
    Z16-940
  • Data Source
    CJP  NDL  NII-ELS 
Page Top